Glioblastoma, the most malignant primary brain tumor, has remained largely refractory to all forms of therapy, and is almost uniformly lethal with a median survival of approximately 15 months. Recently isolated glioblastoma stem cells (GSCs) are thought to be important in the tumor's ability to evade therapy. In addition, we found that the brain tumors they generate retain many of the features of the patient tumors, so they provide a representative tumor model for testing new therapies, something that established cell lines do not. We, and others have developed oncolytic herpes simplex virus (oHSV) vectors that selectively replicate in and kill cancer cells, without harming the surrounding normal tissue or causing disease. The safety of oHSV therapy has been demonstrated in clinical trials for glioblastoma; however efficacy remains anecdotal and needs improvement. The oHSV vectors in clinical trial for glioblastoma contain deletions of the ?34.5 gene, which renders them non-permissive in GSCs. Therefore, we developed a new oHSV, MG18L, deleted for Us3 that is the focus of these studies. The Us3 gene is a serine-threonine kinase that inhibits apoptosis and phosphorylates TSC2 inducing translation. We hypothesize that MG18L selectively replicates in cancer cells because it is unable to promote translation or inhibit virus-induced apoptosis in normal cells, but dysregulated pathways in GSCs makes them permisive. To reduce virus pathogenicity in the brain, MG18L also is mutated in the viral ribonucleotide reductase gene, which targets virus replication to dividing cancer cells. Another feature of Us3-deleted HSV is activation of the PI3K/Akt pathway, genetically altered in the majority of glioblastomas. Therefore, we hypothesize that MG18L will synergize with small molecule inhibitors of the PI3K/Akt pathway, many of which are currently in clinical trial, and kill GSCs and tumors. We recently found that oHSV infection of GSCs targets the ATM pathway and inhibits homologous recombination (HR) DNA repair. Cancer cells deficient in HR are synthetic lethal with poly(ADP-ribose) polymerase (PARP) inhibitors. Thus, we hypothesize that treating GSCs with MG18L will sensitize them to PARP inhibitors inducing a synthetic lethal-like effect. Both of these combination strategies should greatly improve the treatment of glioblastoma and overcome drug resistance. We anticipate that therapies demonstrated in the proposed studies will be translatable to the clinic. Our long-term goal is to develop oHSV vectors that are safe and effective for clinical use and to elucidate combination strategies that will enhance efficacy.